Method and apparatus for vehicle position detection

1. A method for vehicle position detection, comprising:

acquiring a real image of a region in front of a target vehicle, wherein the real image is captured by an image collecting device installed on the target vehicle, and the real image is an image captured by the image collecting device at a current moment under a fixed viewing angle; and

generating a virtual image based on the real image, and determining position information of a side vehicle of the target vehicle based on the virtual image, wherein the virtual image is an image captured by the image collecting device at the current moment under a preset viewing angle, and the preset viewing angle is a viewing angle acquired by rotating the fixed viewing angle by a preset angle in a horizontal direction.

2. The method according claim 1 , wherein the preset angle is a first angle or a second angle, a first viewing angle is a viewing angle acquired by rotating counterclockwise the fixed viewing angle by the first angle in the horizontal direction, and a second viewing angle is a viewing angle acquired by rotating clockwise the fixed viewing angle by the second angle in the horizontal direction; and

the side vehicle of the target vehicle is a left-side vehicle of the target vehicle, in a case that the preset angle is the first angle; and the side vehicle of the target vehicle is a right-side vehicle of the target vehicle, in a case that the preset angle is the second angle.

3. The method according to claim 1 , further comprising:

establishing a coordinate conversion relationship between a real plane coordinate system and a virtual plane coordinate system, wherein the real plane coordinate system is a two-dimensional coordinate system corresponding to an imaging plane under the fixed viewing angle, and the virtual plane coordinate system is a two-dimensional coordinate system corresponding to an imaging plane under the preset viewing angle;

wherein the generating the virtual image based on the real image comprises:

converting the real image to the virtual image according to the coordinate conversion relationship between the real plane coordinate system and the virtual plane coordinate system.

4. The method according to claim 3 , wherein the establishing the coordinate conversion relationship between the real plane coordinate system and the virtual plane coordinate system comprises:

establishing a coordinate conversion relationship between the real plane coordinate system and a real device coordinate system, wherein the real device coordinate system is a three-dimensional coordinate system established for the image collecting device in a case that the image collecting device is corresponding to the fixed viewing angle;

establishing a coordinate conversion relationship between the real device coordinate system and a real world coordinate system, wherein the real world coordinate system is a world coordinate system established in a case that the image collecting device is corresponding to the fixed viewing angle;

establishing a coordinate conversion relationship between the real world coordinate system and a virtual world coordinate system, wherein the virtual world coordinate system is a world coordinate system established in a case that the image collecting device is corresponding to the preset viewing angle;

establishing a coordinate conversion relationship between the virtual world coordinate system and a virtual device coordinate system, wherein the virtual device coordinate system is a three-dimensional coordinate system established for the image collecting device in a case that the image collecting device is corresponding to the preset viewing angle; and

establishing a coordinate conversion relationship between the virtual device coordinate system and the virtual plane coordinate system.

5. The method according to claim 4 , wherein the determining the position information of the side vehicle of the target vehicle based on the virtual image comprises:

detecting a wheel in the virtual image, to acquire a virtual wheel region;

converting the virtual wheel region into the real plane coordinate system according to the coordinate conversion relationship between the real plane coordinate system and the virtual plane coordinate system, to acquire a converted real wheel region; and

converting the converted real wheel region into the real world coordinate system according to a coordinate conversion relationship between the real plane coordinate system and the real world coordinate system, to acquire position information of the wheel of the side vehicle of the target vehicle.

6. The method according to claim 5 , further comprising:

learning a wheel feature in a large quantity of vehicle sample images, to acquire a wheel recognition model used for recognizing the wheel;

wherein the detecting the wheel in the virtual image comprises:

detecting the wheel in the virtual image by using the wheel recognition model.

7. The method according to claim 5 , further comprising:

performing optic flow tracking on the wheel of the side vehicle of the target vehicle, to acquire a real wheel region at the current moment;

wherein after acquiring the converted real wheel region, the method further comprises:

fusing the converted real wheel region and the real wheel region acquired by the optic flow tracking, to acquire a fused real wheel region.

8. The method according to claim 7 , wherein the acquiring the real wheel region at the current moment comprises:

performing edge supplement on an optic flow tracking image at the current moment; and

acquiring the real wheel region of the wheel of the side vehicle based on the supplemented optic flow tracking image.

9. The method according to claim 7 , wherein after acquiring the fused real wheel region, the method further comprises:

acquiring a final real wheel region corresponding to each of one or more sequential history moments before the current moment; and

determining whether there is a non-wheel region in the fused real wheel region based on the final real wheel region corresponding to each of the one or more sequential history moments, and removing the non-wheel region in case that there is the non-wheel region in the fused real wheel region.

10. The method according to claim 1 , wherein:

the real image does not include a feature effective for recognizing the side vehicle, and the virtual image includes the feature effective for recognizing the side vehicle; and

the preset angle is adjusted in real time for capturing the feature effective for recognizing the side vehicle.

11. An apparatus for vehicle position detection comprising a processor, a memory for storing program instructions, and a system bus, wherein the processor and the memory are connected via the system bus, and the processor executes the program instructions to:

acquire a real image of a region in front of a target vehicle, wherein the real image is captured by an image collecting device installed on the target vehicle, and the real image is an image captured by the image collecting device at a current moment under a fixed viewing angle;

generate a virtual image based on the real image, wherein the virtual image is an image captured by the image collecting device at the current moment under a preset viewing angle, and the preset viewing angle is a viewing angle acquired by rotating the fixed viewing angle by a preset angle in a horizontal direction; and

determine position information of a side vehicle of the target vehicle based on the virtual image.

12. The apparatus according to claim 11 , wherein the preset angle is a first angle or a second angle, a first viewing angle is a viewing angle acquired by rotating counterclockwise the fixed viewing angle by the first angle in the horizontal direction, and a second viewing angle is a viewing angle acquired by rotating clockwise the fixed viewing angle by the second angle in the horizontal direction; and

the side vehicle of the target vehicle is a left-side vehicle of the target vehicle, in a case that the preset angle is the first angle; and the side vehicle of the target vehicle is a right-side vehicle of the target vehicle, in a case that the preset angle is the second angle.

13. The apparatus according to claim 11 , wherein the processor executes the program instructions further to:

establish a coordinate conversion relationship between a real plane coordinate system and a virtual plane coordinate system, wherein the real plane coordinate system is a two-dimensional coordinate system corresponding to an imaging plane under the fixed viewing angle, and the virtual plane coordinate system is a two-dimensional coordinate system corresponding to an imaging plane under the preset viewing angle; and

convert the real image to the virtual image according to the coordinate conversion relationship between the real plane coordinate system and the virtual plane coordinate system.

14. The apparatus according to claim 13 , wherein the processor executes the program instructions further to:

establish a coordinate conversion relationship between the real plane coordinate system and a real device coordinate system, wherein the real device coordinate system is a three-dimensional coordinate system established for the image collecting device in a case that the image collecting device is corresponding to the fixed viewing angle;

establish a coordinate conversion relationship between the real device coordinate system and a real world coordinate system, wherein the real world coordinate system is a world coordinate system established in a case that the image collecting device is corresponding to the fixed viewing angle;

establish a coordinate conversion relationship between the real world coordinate system and a virtual world coordinate system, wherein the virtual world coordinate system is a world coordinate system established in a case that the image collecting device is corresponding to the preset viewing angle;

establish a coordinate conversion relationship between the virtual world coordinate system and a virtual device coordinate system, wherein the virtual device coordinate system is a three-dimensional coordinate system established for the image collecting device in a case that the image collecting device is corresponding to the preset viewing angle; and

establish a coordinate conversion relationship between the virtual device coordinate system and the virtual plane coordinate system.

15. The apparatus according to claim 14 , wherein the processor executes the program instructions further to:

detect a wheel in the virtual image, to acquire a virtual wheel region;

convert the virtual wheel region into the real plane coordinate system according to the coordinate conversion relationship between the real plane coordinate system and the virtual plane coordinate system, to acquire a converted real wheel region; and

convert the converted real wheel region into the real world coordinate system according to a coordinate conversion relationship between the real plane coordinate system and the real world coordinate system, to acquire position information of the wheel of the side vehicle of the target vehicle.

16. The apparatus according to claim 15 , wherein the processor executes the program instructions further to:

learn a wheel feature in a large quantity of vehicle sample images, to acquire a wheel recognition model used for recognizing the wheel; and

detect the wheel in the virtual image by using the wheel recognition model.

17. The apparatus according to claim 15 , wherein the processor executes the program instructions further to:

perform optic flow tracking on the wheel of the side vehicle of the target vehicle, to acquire a real wheel region at the current moment; and

fuse, after the converted real wheel region is acquired, the converted real wheel region and the real wheel region acquired by the optic flow tracking, to acquire a fused real wheel region.

18. The apparatus according to claim 17 , wherein the processor executes the program instructions further to:

perform the optic flow tracking on the wheel of the side vehicle of the target vehicle;

perform edge supplement on an optic flow tracking image at the current moment; and

acquire the real wheel region of the wheel of the side vehicle based on the supplemented optic flow tracking image.

19. The apparatus according to claim 17 , wherein the processor executes the program instructions further to:

acquire, after the fused real wheel region is acquired, a final real wheel region corresponding to each of one or more sequential history moments before the current moment; and

determine whether there is a non-wheel region in the fused real wheel region based on the final real wheel region corresponding to each of the one or more sequential history moments, and remove the non-wheel region in case that there is the non-wheel region in the fused real wheel region.

20. A non-transitory computer-readable storage medium storing instructions, wherein the instructions when being executed on a terminal device cause the terminal device to:

acquire a real image of a region in front of a target vehicle, wherein the real image is captured by an image collecting device installed on the target vehicle, and the real image is an image captured by the image collecting device at a current moment under a fixed viewing angle; and

generate a virtual image based on the real image, and determine position information of a side vehicle of the target vehicle based on the virtual image, wherein the virtual image is an image captured by the image collecting device at the current moment under a preset viewing angle, and the preset viewing angle is a viewing angle acquired by rotating the fixed viewing angle by a preset angle in a horizontal direction.